Nvidia Management Explains: Where Does the Market-Shaking 70% Revenue Guidance Come From?

Nvidia rarely provides multi-year revenue outlooks to the market, and this move immediately drew widespread attention.

In a report released on Sept. 2, JPMorgan said that based on recent discussions with Nvidia's Vice President of Investor Relations and Strategic Finance, Toshiya Hari, this 70% year-over-year growth framework does not stem from a single driver, but is built on the basis of accelerating demand across hyperscalers, emerging cloud service providers, AI labs, sovereign AI, and enterprise customers. More critically, management explicitly stated that without supply constraints, business growth could potentially double—the core variable currently limiting growth is supply, not demand.

The direct motivation for Nvidia proactively disclosing forward-looking guidance for fiscal 2028 is to bridge the significant gap between market consensus and the company's internal assessment. Management believes that if this gap persists, it could pose substantial challenges to the planning of supply chain partners. This statement implies that the 70% growth target is a well-supported "comfort zone" in management's view, rather than an aggressive forecast, and the implied upside far exceeds the market's current pricing.

Meanwhile, Nvidia has sent important signals across multiple dimensions including customer structure, inference business share, supply bottlenecks, and financing arrangements, further outlining the medium-term growth picture for this AI chip giant. JPMorgan maintains an Overweight rating on Nvidia with a price target of $320, implying approximately 43% upside from the current stock price of $224.41 (Sept. 2 close).

Supply, Not Demand, Is the Real Ceiling on Growth

The report said that in discussions with Nvidia's Vice President of Investor Relations and Strategic Finance, Toshiya Hari, JPMorgan found that Nvidia management expressed clear confidence in the 70% growth framework, but also delineated its boundaries: this is a supply-constrained number, not a demand-constrained one.

According to the JPMorgan report, Toshiya Hari pointed out in the discussions that if supply constraints were excluded, Nvidia's business growth could potentially more than double year-over-year. This statement directly reveals the conservative nature of the current guidance—70% is an expectation under realistic supply conditions, not the upper limit of demand the company could reach.

Management also specifically explained that the decision to provide a multi-year outlook partly stems from a "meaningful gap" between market consensus and the company's internal observations. If not proactively disclosed, this information asymmetry could lead to mismatches in capacity planning by supply chain partners, which in turn could constrain Nvidia's own delivery capabilities. In other words, this forward-looking guidance is both a signal to investors and proactive management of the supply chain ecosystem.

Advanced Wafers and Memory: Two Core Bottlenecks in the Supply Chain

Regarding the specific composition of supply constraints, Toshiya Hari named two most critical materials: advanced wafers and memory, listing them as the two highest-weighted items in Nvidia's bill of materials (BOM).

According to the JPMorgan report, Nvidia is currently in deep communication with TSMC and three memory suppliers—Micron, SK Hynix, and Samsung—with core discussions centered on improving supply availability.

The bank believes that Toshiya Hari's statement indicates Nvidia's supply chain management has entered a phase of high-intensity proactive coordination, rather than passively waiting for capacity release. The pace of alleviating these bottlenecks will directly determine whether Nvidia can achieve growth exceeding the 70% baseline in FY28. The room for supply chain improvement is the room for performance elasticity.

Inference Business Share Continues to Expand, but Platform Versatility Makes Quantification Difficult

The revenue structure between inference and training is a long-standing market concern. Toshiya Hari provided the clearest directional judgment to date: about 18 months ago, training and inference revenue shares were roughly equal; currently, inference has surpassed training, and this trend is expected to continue.

However, the report said management also pointed out that precisely quantifying the split between the two is inherently difficult. The reason lies in the high fungibility of Nvidia's platform—taking Grace Blackwell products as an example, customers can first use them for training workloads and later repurpose the same hardware assets for inference tasks. This flexibility reflects the competitiveness of Nvidia's platform, but it also limits external analysis of the revenue structure breakdown.

Customer Structure Continues to Diversify, Emerging Cloud Providers Contribute Over 50%

Nvidia's revenue sources are spreading from hyperscalers to a broader ecosystem. According to Toshiya Hari, OpenAI and Anthropic, the two leading frontier model builders, currently account for about 20% of Nvidia's business on an end-consumption basis, and this proportion is expected to rise to about 25% by FY28.

It is worth noting that the above figures reflect the share at the end-consumption level, not Nvidia's direct customer structure—Nvidia typically sells compute to hyperscalers or emerging cloud service providers, who then sell compute capacity to model builders.

At the direct customer level, the contribution of emerging cloud service providers (neoclouds) can no longer be ignored. According to the JPMorgan report, neoclouds currently account for more than 50% of Nvidia's ACIE (Accelerated Computing and AI Infrastructure Ecosystem) business, indicating that Nvidia's growth engine no longer relies solely on a few top cloud providers, but is jointly driven by a broader ecosystem of compute construction and leasing.

Open Source vs. Closed Source: Nvidia's Answer Is "Both Are Needed"

Amid market debates over the merits of open-source versus closed-source large language models (LLMs), Nvidia management gave a clear stance: the two are not an either-or competitive relationship; the continued evolution of AI requires the synergistic development of open-source and closed-source models.

Toshiya Hari said that Nvidia internally uses closed-source models such as OpenAI and Claude extensively, while in critical task scenarios like chip design, it adopts a combination of closed-source and open-source approaches. Management's core logic is that as long as model builders can achieve commercialization and continuously improve their economic models, demand will continue to transmit to chip suppliers like Nvidia.

Additionally, management mentioned that model builders' gross margins (GM) appear to be improving, partly due to the per-token cost reduction across generations driven by Nvidia's platform. The improvement in model economics forms a positive feedback loop for demand for Nvidia chips.

Financing Arrangements Aim to Support Forward Demand, Management Rebuts "Circular Financing" Concerns

Nvidia's recent introduction of multiple financing arrangements has drawn market attention, and management provided a systematic explanation during these discussions.

According to the JPMorgan report, the relevant arrangements include: revenue-sharing agreements signed with certain neoclouds, the PORTS-Pike data center campus project, and a $500 billion private capital financing platform established together with institutions such as Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR.

In terms of revenue-sharing agreements, Nvidia's mechanism is to set a floor price for compute leasing and share upside returns when market leasing prices exceed the baseline, thereby creating recurring revenue options beyond core hardware sales.

In response to external concerns about "circular financing," management clearly stated that the relevant financing arrangements are moderate in scale, capped, and supported by strong underlying demand, ecosystem returns, and the creditworthiness of ultimate compute buyers. Nvidia positions these financing tools as a means to support forward demand for AI infrastructure, not as financial leverage operations.